#!/usr/bin/env python
# coding: utf-8

# ## Titanic with Keras

# In[1]:


import numpy as np
import pandas as pd

import keras
from keras.models import Sequential
from keras.layers.normalization import BatchNormalization
from keras.layers.core import Flatten, Dense, Activation, Dropout


# ## setup

# In[24]:


DATA_HOME_DIR = "/home/tsu-nera/repo/kaggle/input/titanic/"
row_data = pd.read_csv(DATA_HOME_DIR + 'train.csv', index_col=0)
test_data = pd.read_csv(DATA_HOME_DIR + 'test.csv', index_col=0)


# In[25]:


test_ind = test_data.index

data= row_data[['Survived','Pclass','Sex','Age','SibSp','Parch','Fare']].dropna()
t_data = test_data[['Pclass','Sex','Age','SibSp','Parch','Fare']]


# In[71]:


Pclass = pd.get_dummies(data['Pclass'])
Pclass.columns=['1st','2nd','3rd']
Pclass = Pclass.drop('1st',axis=1)

tPclass = pd.get_dummies(t_data['Pclass'])
tPclass.columns=['1st','2nd','3rd']
tPclass = tPclass.drop('1st',axis=1)

Sex = pd.get_dummies(data['Sex'])
Sex = Sex.drop('male',axis=1)

tSex = pd.get_dummies(t_data['Sex'])
tSex = tSex.drop('male',axis=1)

def adult_kids(age):
    if age > 15:
        if age > 60:
            return  'older'
        else :
            return   'adult'
    else :
        if age < 16:
            return   'kids'
        else :
            return np.nan

data['adult_kids'] = data['Age'].apply(adult_kids)
Age_cat = pd.get_dummies(data['adult_kids']).drop('adult',axis=1)
t_data['adult_kids'] = t_data['Age'].apply(adult_kids)
tAge_cat = pd.get_dummies(t_data['adult_kids']).drop('adult',axis=1)

Data_tmp = data[['Survived','SibSp','Parch','Fare']]

Merge_data = pd.merge(Data_tmp,Pclass,right_index=True,left_index=True)
Merge_data = pd.merge(Merge_data,Sex,right_index=True,left_index=True)
Merge_data = pd.merge(Merge_data,Age_cat,right_index=True,left_index=True)
y = Merge_data['Survived'].values
x = Merge_data.drop('Survived',axis=1).values

Data_tmp = t_data[['SibSp','Parch','Fare']]

Merge_data = pd.merge(Data_tmp,tPclass,right_index=True,left_index=True)
Merge_data = pd.merge(Merge_data,tSex,right_index=True,left_index=True)
Merge_data = pd.merge(Merge_data,tAge_cat,right_index=True,left_index=True)
tx = Merge_data

(x.shape, y.shape, tx.shape)


# ## Build Model

# In[105]:


# create model
model = Sequential()
model.add(Dense(64, input_shape=(8,)))

for i in range(0, 8):
    model.add(Dense(units=64))
    model.add(Activation('relu'))

model.add(Dense(units=1))
model.add(Activation('linear'))

model.compile(loss='mean_squared_error', optimizer='rmsprop')


# ## Training

# In[102]:


model.fit(x, y, epochs=300)


# ## Testing

# In[103]:


p_survived = model.predict_classes(tx.values)


# In[104]:


submission = pd.DataFrame()
submission['PassengerId'] = test_ind
submission['Survived'] = p_survived


# In[96]:


submission.to_csv('submission.csv', index=False)


# In[ ]: